SANDIA REPORT

SAND2008-xxxx
Unlimited Release
Printed December, 2008

Domain model and implementation of

cell topology data

Pavel Bochev, Harold Edwards, Roger Pawlowski, Denis Ridzal

Prepared by
Sandia National Laboratories
Albuquerque, New Mexico 87185 and Livermore, California 94550

Sandia is a multiprogram laboratory operated by Sandia Corporation,

a Lockheed Martin Company, for the United States Department of Energy’s
National Nuclear Security Administration under Contract DE-AC04-94-AL85000.

Approved for public release; further dissemination unlimited.

Issued by Sandia National Laboratories, operated for the United States Department of Energy by Sandia
Corporation.

NOTICE: This report was prepared as an account of work sponsored by an agency of the United States
Government. Neither the United States Government, nor any agency thereof, nor any of their employees,
nor any of their contractors, subcontractors, or their employees, make any warranty, express or implied,
or assume any legal liability or responsibility for the accuracy, completeness, or usefulness of any infor-
mation, apparatus, product, or process disclosed, or represent that its use would not infringe privately
owned rights. Reference herein to any specific commercial product, process, or service by trade name,
trademark, manufacturer, or otherwise, does not necessarily constitute or imply its endorsement, recom-
mendation, or favoring by the United States Government, any agency thereof, or any of their contractors
or subcontractors. The views and opinions expressed herein do not necessarily state or reflect those of
the United States Government, any agency thereof, or any of their contractors.

Printed in the United States of America. This report has been reproduced directly from the best available
copy.

Available to DOE and DOE contractors from

U.S. Department of Energy
Office of Scientific and Technical Information
P.O. Box 62
Oak Ridge, TN 37831

Telephone: (865) 576-8401

Facsimile: (865) 576-5728
E-Mail: reports@adonis.osti.gov
Online ordering: http://www.osti.gov/bridge

Available to the public from

U.S. Department of Commerce
National Technical Information Service
5285 Port Royal Rd
Springfield, VA 22161

Telephone: (800) 553-6847

Facsimile: (703) 605-6900
E-Mail: orders@ntis.fedworld.gov
Online ordering: http://www.ntis.gov/help/ordermethods.asp?loc=7-4-0#online

NT OF E
ME N
RT
ER
A
DEP

GY

• •
IC A
U NIT

ER
ED

ST M
A TES OF A

2
 SAND2008-xxxx
Unlimited Release
Printed December, 2008

Domain model and implementation of cell

topology data

Pavel Bochev
Applied Mathematics and Application Dept. 1414, MS-1320
Sandia National Laboratories
Albuquerque, NM 87185-1320
pbboche@sandia.gov

Harold C. Edwards
Computational Thermal and Fluid Mechanics Dept. 1542, MS-????
Sandia National Laboratories
Albuquerque, NM 87185-????
hcedwar@sandia.gov

Roger Pawlowski
Applied Mathematics and Application Dept. 1414, MS-1320
Sandia National Laboratories
Albuquerque, NM 87185-1320
rppawlo@sandia.gov

Denis Ridzal
Optimization and UQ Dept. 1411, MS-1320
Sandia National Laboratories
Albuquerque, NM 87185-1320
dridzal@sandia.gov

Abstract

Numerical solution of Partial Differential Equations (PDEs) by finite ele-

ment, finite volume and finite difference methods requires partitioning of the

3
computational domain into subdomains commonly referred to as elements, cells,
or zones. This document describes a domain model for subdomain data in terms
of standardized cell topologies that can be shared across a wide range of applica-
tion codes for mesh-based solution of PDEs. The domain model is implemented
in a Trilinos package, called Shards, which provides collection of default topolo-
gies and a mechanism for creating user-defined custom cell topologies. The use
of this package and its features are briefly explained.

4
 Acknowledgment

R. Bartlett, Greg Sjaardema and ????? provided useful insight and comments during
the work on this project and the preparation of this document. Their help is greatly
appreciated.

5
6
Contents
1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11
2 Domain model for cell topology data . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12
2.1 Geometry of polytopes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12
2.2 Computational cells . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13
3 Implementation in Shards . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18
3.1 Cell topology . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18
3.2 Canonical cell topologies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18
4 Using cell topologies. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19
4.1 Custom cell topologies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20

Appendix
A Cell Topology in Shards . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21
A.1 Line Topology . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21
A.2 Triangle . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21
A.3 Quadrilateral . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21
A.4 Pentagon . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21
A.5 Hexagon . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22
A.6 Tetrahedron . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22
A.7 Pyramid . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23
A.8 Wedge . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23
A.9 Hexahedron . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24

7
Figures
A.1 (a) Base line (line<2>) and (b) extended line (line<3>) topology in
Shards. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21
A.2 (a) Base triangle (Triangle<>), (b) four-node extended triangle (Triangle<4>),
and (c) six-node extended triangle (Triangle<6>) topology in Shards. 22
A.3 Triangle edge (1-cell) numbering in Shards. . . . . . . . . . . . . . . . . . . . . . . . 23
A.4 (a) Base quadrilateral (Quadrilateral<4>), (b) eight-node extended
quadrilateral (Quadrilateral<8>), and (c) nine-node extended quadri-
lateral (Quadrilateral<9>) topology in Shards. . . . . . . . . . . . . . . . . . . 23
A.5 Quadrilateral edge (1-cell) numbering in Shards. . . . . . . . . . . . . . . . . . . . 24
A.6 (a) Base pentagon (Pentagon<>) topology and (b) edge (1-cell) num-
bering in Shards. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24
A.7 (a) Base hexagon (Hexagon<>) topology and (b) edge (1-cell) number-
ing in Shards. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25
A.8 (a) Base tetrahedron (Tetrahedron<>), (b) eight-node extended tetra-
hedron (Tetrahedron<8>), and (e) ten-node extended tetrahedron
(Tetrahedron<10>) topology in Shards. . . . . . . . . . . . . . . . . . . . . . . . . . 25
A.9 Tetrahedron edge (1-cell) numbering in Shards. . . . . . . . . . . . . . . . . . . . . 26
A.10 Tetrahedron face (2-cell) numbering in Shards. . . . . . . . . . . . . . . . . . . . . 26
A.11 (a) Base pyramid (Pyramid<>), (b) thirteen-node extended pyramid
(Pyramid<13>), and (c) fourteen-node extended pyramid (Pyramid<14>)
topology in Shards. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 27
A.12 Pyramid edge (1-cell) numbering in Shards. . . . . . . . . . . . . . . . . . . . . . . . 27
A.13 Pyramid face (2-cell) numbering in Shards. . . . . . . . . . . . . . . . . . . . . . . . 28
A.14 (a) Base wedge (Wedge<>), (b) fifteen-node extended wedge (Wedge<15>),
and (c) eighteen-node extended wedge (Wedge<18>) topology in Shards. 28
A.15 Wedge edge (1-cell) numbering in Shards. . . . . . . . . . . . . . . . . . . . . . . . . 29
A.16 Wedge face (2-cell) numbering in Shards. . . . . . . . . . . . . . . . . . . . . . . . . . 29
A.17 (a) Base hexahedron(Hexahedron<>), (b) twenty-node extended hex-
ahedron (Hexahedron<20>), and (c) twenty seven-node extended hex-
ahedron (Hexahedron<27>) topology in Shards. . . . . . . . . . . . . . . . . . . . 30
A.18 Hexahedron edge (1-cell) numbering in Shards. . . . . . . . . . . . . . . . . . . . . 30
A.19 Hexahedron face (2-cell) numbering in Shards. . . . . . . . . . . . . . . . . . . . . 31

8
Tables

9
10
 1 Introduction

The use of grids to partition the computational domain into a union of smaller sub-
domains is a hallmark of a large number of methods for the approximate numerical
solution of PDEs; see [5] package Shards is discussed in Section 3. This section also
contains a summary of the canonical cell topologies provided in Shards. The report
concludes with several use cases presented in Section 4. Among other things these
examples cover instantiation of custom user-defined cell topologies.

11
 2 Domain model for cell topology data

2.1 Geometry of polytopes

A computational domain Ω ⊂ Rd can be partitioned into subdomains (cells) with

arbitrary shapes in many possible ways. However, practical considerations such as
having simple and effective means of encoding the topology of these subdomains in
computer programs, restrict the range of cell shapes used in mesh-based methods to
polytopes.

Definition 1. A d-dimensional polytope, d > 0 is a closed bounded region in Rd

defined by an intersection of a finite number of hyperplanes. A 0-dimensional, or
trivial polytope, is a point. 2

From Definition 1 it follows that for d > 0 a d-dimensional polytope always has
a non-empty1 interior. Furthermore, this definition implies that any d-dimensional
polytope Pd can be described by a set of ` linear constraints

Pd = {x ∈ Rd | Ax ≤ b} (2.1)

where A is a real ` × d matrix and b is a real `-dimensional vector; see [7].

Example 1. The canonical 2-simplex σ2 (a triangle) is a two-dimensional polytope

that is the convex span of {(0, 0), (1, 0), (0, 1)}. This polytope is described by the
following set of inequalities:

 x ≥0
(x1 , x2 ) ∈ σ2 ⇔ y ≥0
x+y ≤1


It follows that σ2 = {x ∈ R2 | Ax ≤ b} with

−1
   
0 0
A =  0 −1  and b= 0 .
1 1 1
2

Any d-dimensional polytope Pd can be embedded in a higher-dimensional space

n
R , n > d in an obvious manner. Specifically, assuming that Pd is defined by (2.1),
its embedding in Rn is given by

Pbd = {x ∈ Rn | Ax
b ≤ b ; xd+1 = . . . = xn = 0} ; (2.2)
1
This is because a region is an open and connected set.

12
where Ab = [A 0] is ` × n matrix. Referring to illustration in Example 1 we see that
that the embedding of the canonical 2-simplex in R3 corresponds to
−1
 
0 0
b =  0 −1 0  and x3 = 0 .
A (2.3)
1 1 0

The possibility for a polytope to be embedded in a higher-dimensional space mo-

tivates the following definition.
Definition 2. Let Pk denote a k-dimensional polytope embedded in Rd , d ≥ k. The
number k is called topological dimension of Pk and d is its physical dimension. 2

The physical dimension of a polytope is always greater than or equal to its topo-
logical dimension, i.e., a non-trivial polytope cannot be realized in a physical space
of lower dimension than its topological dimension. Note that the topological dimen-
sion is the smallest physical dimension in which a non-trivial polytope has non-empty
interior.

2.2 Computational cells

The notion of a polytope is defined for any dimension n. Because computational

domains that arise in the numerical solution of PDEs are typically subsets of Rd ,
d = 1, 2, 3 this generality is not necessary for the subsequent discussion. For this
reason we limit the space dimension to 3 and make the following definition.
Definition 3. A standard d-cell, or simply a d-cell is a d-dimensional polytope, d =
0, 1, 2, 3. 2

We see that d-cells correspond to familiar geometrical objects:

ˆ 0-cell is a point (trivial polytope);

ˆ 1-cell is a line (1-dimensional polytope);
ˆ 2-cell is a polygon (2-dimensional polytope);
ˆ 3-cell is a polyhedron (3-dimensional polytope).

By virtue of being k-dimensional polytopes, k-cells inherit all properties of general

polytopes. In particular, a k-cell Pk can be embedded in Rd as long as d ≥ k. Also,
a non-trivial k-cell has non-empty interior if and only if its topological dimension
coincides with its physical dimension, i.e., when k = d. In other words, any k-cell
embedded in a higher-dimensional space has empty interior relative to that space.
This necessitates an alternative definition of the boundary operator for embedded
cells that recovers their “true” boundary.

13
Definition 4. Assume that Pk is a k-cell in defined by (2.1) with physical dimension
d ≥ k. The topological boundary ∂τ Pk of Pk is defined to be the following set:
" ` #
[ \
∂τ Pk = {x ∈ Rk | Ai x = bi } Pk . (2.4)
i=1

where Ai is the ith row of A. The physical boundary ∂Pk of Pk is defined in the usual
manner2 as
∂Pk = Pk ∩ (Rd \ Pk ) (2.5)
2

The topological boundary of a k-cell is always a union of m-cells with smaller

topological dimensions, i.e., with 0 ≤ m < k. On the other hand, whenever the
topological dimension of a k-cell is less than its physical dimensions we have that
∂Pk = Pk . Therefore, the topological and the physical boundary of a k-cell coincide
if and only if Pk has the same topological and physical dimensions:

∂τ Pk = ∂Pk ⇔ k = d.

Example 2. Consider the 2-simplex from Example 1 and assume that it is embedded
in R3 . Then, the physical boundary of σ2 is the 2-simplex itself

∂σ2 = σ2 ,

because σ2 is a closed set with an empty interior in R3 . On the other hand, the
topological boundary of σ2 is the set

∂τ σ2 = {x ∈ R3 | Ax
b = b ; x3 = 0} ∩ σ2

where Ab is the matrix defined in (2.3) and b is the vector from Example 1. This set
consists of the three lines enclosing the 2-simplex and does not include σ2 itself.

The notion of topological boundary makes it possible to introduce the useful

concept of a subcell.

Definition 5. For 0 ≤ m < k the m-cell Sm is called m-subcell of Pk if and only if it

belongs to the topological boundary of Pk , i.e., Sm ⊂ ∂τ Pk . If m = k the only k-subcell
of Pk is defined to be Pk itself. The k-cell Pk is called parent cell of its m-subcell Sm .

For m < k the set of all m-subcells of Pk is given by

[ [
Σm (Pk ) = ··· Sm . (2.6)
Sk−1 ∈∂τ Pk Sm ∈∂τ Sm+1

2
We remind that, by definition, Pk is a closed set, i.e., Pk = Pk .

14
 We also have the obvious identities
(
∂τ Pk if k = d
∂τ Pk = Σ0 (Pk ) ∪ · · · ∪ Σk−1 (Pk ) and ∂Pk =
Σk (Pk ) if k < d

Here, Σ0 (Pk ) is the set of all 0-subcells, Σ1 (Pk ) - the set of all 1-subcells and so
on. Of course, Σk (Pk ) = Pk . While dim Σk (Pk ) = 1 for any standard cell, the number
of lower-dimensional subcells varies with the cell shape. Mathematically, this shape
is encoded by the notion of a cell topology.

Definition 6. An enumeration of the set Σ0 (Pk ), i.e., a surjective mapping from N

to Σ0 (Pk ), is called cell topology of Pk and denoted by τ (Pk ).

In words, cell topology is a map that assigns a unique integer ir to every 0-subcell
in Σ0 (Pk ):
τ : Σ0 (Pk ) 3 S0r 7→ ir ∈ N such that ir 6= it for r 6= t .
Given a k-cell Pk the number of all possible cell topologies equals (dim Σ0 (Pk ))!.
However, once a cell topology τ (Pk ) is selected, it induces a unique topology τ (Sm )
on every m-subcell of Pk .

2.2.1 Shell cells

The case of a d-cell embedded in Rd+1 deserves special attention because of the
possibility to endow such cells by an alternative shell topology. We first define the
notion of a shell cell.

Definition 7. Let Pd denote a standard d-cell embedded in Rd+1 , 0 < d < 3. The
shell extension of Pd , denoted by [Pd ], is defined to be the set
d−1
!
[ [
[Pd ] = Σm (Pd ) {Pd , Pd } . (2.7)
m=0

In particular, for a shell extension of Pd we have that

(
Σm (Pd ) for 0 < m < d
Σm ([Pd ]) =
{Pd , Pd } for m = d

Simply put, extension of a d-cell to a shell d-cell doubles the number of its d-
subcells. A shell cell [Pd ] can be endowed with a shell topology in a natural way by
extending the topology of its standard prototype Pd .

15
Definition 8. Let Pd denote a d-cell embedded in Rd+1 , 0 < d < 3 and [Pd ] is its
shell extension. Assume that Pd is endowed with cell topology τ (Pd ). The induced
shell topology on [Pd ] is given by
(
τ (Σm (Pd )) for 0 < m < d
τ (Σm ([Pd ])) =
{τ (Pd ), Πτ (Pd )} for m = d

where Π is left or right cyclic permutation. 2

This definition means that all lower-dimensional subcells of a shell cell retain the
topology of the standard prototype Pd , one of the d-subcells inherits the topology of
Pd and the other is endowed with a version of that topology obtained by a left or
right cyclic permutation of τ

2.2.2 Classical nomenclature

Definitions 4–8 provide “standard” terminology that makes mathematically precise

key notions needed to represent cells and query their topologies in most mesh-based
methods for numerical PDEs. Because an alternative set of “classical” definitions has
been in widespread use in legacy codes and frameworks it is important to define the
“classical” notions in terms of the standard definitions. This also offers the added
benefit of standardizing the classical definitions by mapping them to mathematically
precise set of standard definitions.

Definition 9.

ˆ A node is any 0-cell;

ˆ A vertex is a 0-cell that is part of the topological boundary of a d-cell for d ≥ 1;

ˆ An edge is a 1-cell that is part of the topological boundary of a d-cell for d ≥ 2;

ˆ A face is a 2-cell that is part of the topological boundary of a d-cell for d = 3;

ˆ A side3 is a d − 1-cell that is part of the boundary of a d-cell for d > 0.

ˆ An element is a d-cell whose physical dimension equals its topological dimen-

sion.

ˆ A shell element is a shell d-cell, i.e., a shell extension of a standard d-cell

embedded in Rd+1 .

ˆ A beam element is a d-cell embedded in Rd+2 .

3
The d − 1-subcells are sometimes also called facets; see [7].

16
 While the concept of an m-subcell is quite similar to that of a vertex, edge,
face or a side, there are some important distinctions between classical and standard
definitions. Most notably, the notion of an m-subcell allows the subcell to have the
same topological dimension as its parent cell. This proved convenient in the definition
of shell extensions of standard cells. In contrast, the “classical” notions define vertex,
edge, face and side as necessarily being parts of the topological boundary of some
parent cell, i.e., their topological dimensions are always less than the topological
dimension of that parent cell.

For a simple example consider a 1-cell P1 , i.e., a line. According to Definition 5

P1 has two 0-subcells (the endpoints) and one 1-subcell - the line P1 itself. Switching
to the classical notions we can say that P1 has two vertices but we can’t say that
it has one edge because definition of the latter requires a parent cell of topological
dimension at least two.

Open question. How to reconcile with shell elements which are allowed to have
faces?

17
 3 Implementation in Shards

3.1 Cell topology

subcell homogeneity: The set of d-subcells is homogeneous if every d-subcell has the
same topology. Note that this is based on the actual and not the base topology of
the d-cell. As a result, this definition includes a case where all d-subcells may have
the same base topology, e.g., Line¡¿, but their individual topologies may be different,
e.g., a triangle with edges given by:

Line¡¿, Line¡3¿, Line¡¿

This configuration could come up in hp-methods.

base topology

cell topology

3.2 Canonical cell topologies

18
 4 Using cell topologies

4.1 Custom cell topologies

19
 References
[1] P. Bochev and M. Hyman. Principles of compatible discretizations. In D. Arnold,
P. Bochev, R. Lehoucq, R. Nicolaides, and M. Shashkov, editors, Compatible dis-
cretizations, Proceedings of IMA Hot Topics workshop on Compatible discretiza-
tions, volume IMA 142, pages 89–120. Springer Verlag, 2006.

[2] F. Brezzi, K. Lipnikov, M. Shashkov and V. Simoncini, A new discretization

methodology for diffusion problems on generalized polyhedral meshes Computer
Methods in Applied Mechanics and Engineering, Vol. 196, Issues 37-40, 1 August
2007, pp. 3682-3692

[3] F. Brezzi, K. Lipnikov, and V. Simoncini, A family of mimetic finite difference

methods on polygonal and polyhedral meshes M3AS: Mathematical Models and
Methods in Applied Sciences, v.15 n.10 (2005) pp. 1533-1552.

[4] P. Ciarlet. The finite element method for elliptic problems. SIAM Classics in
Applied Mathematics. SIAM, 2002.

[5] M. A. Heroux, R. A. Bartlett, V. E. Howle, R. J. Hoekstra, J. J. Hu, T. G. Kolda,

R. B. Lehoucq, K. R. Long, R. P. Pawlowski, E. T. Phipps, A. G. Salinger, H. K.
Thornquist, R. S. Tuminaro, J. M. Willenbring, A. Williams, and K. S. Stanley.
An overview of the Trilinos project. ACM Trans. Math. Softw., 31(3):397–423,
2005.

[6] M. Shashkov. Conservative finite difference methods on general grids. CRC

Press, Boca Raton, FL, 1996.

[7] E. W. Weisstein, ”Polytope.” From MathWorld–A Wolfram Web Resource.

http://mathworld.wolfram.com/Polytope.html

20
 A Cell Topology in Shards

This appendix describes the topology of the cell types provided in Shards. Many of
the cell types have a base topology that includes a minimum number of nodes and
one or more extended topologies with an increasing number of nodes.

A.1 Line Topology

A line is an object of dimension 1 with two vertices. Both base and extended topolo-
gies exist for a line in Shards as shown in Figure A.1.

0 1 0 2 1

(a) (b)

Figure A.1: (a) Base line (line<2>) and (b) extended line (line<3>) topology in
Shards.

A.2 Triangle

A triangle is an object of dimension two with three vertices and three edges. The
base and extended topologies available in Shards are shown in Figure ??. The edge
numbering is shown in Figure A.3.

A.3 Quadrilateral

A quadrilateral is an object of dimension two with four vertices and four edges. The
base and extended topologies available in Shards are shown in Figure A.4. The edge
numbering is shown in Figure A.5.

A.4 Pentagon

A pentagon is an object of dimension two with five vertices and five edges. Only a
base topology is provided for a pentagon in Shards, which is shown in Figure A.6
along with the edge numbering.

21
 2 2

0 1 0 1

(a) (b)

5 4

0 1
3

(c)

Figure A.2: (a) Base triangle (Triangle<>), (b) four-node extended triangle
(Triangle<4>), and (c) six-node extended triangle (Triangle<6>) topology in
Shards.

A.5 Hexagon

A hexagon is an object of dimension two with six vertices and six edges. Only a base
topology is provided for a hexagon in Shards, which is shown in Figure ?? along with
the edge numbering.

A.6 Tetrahedron

A tetrahedron is an object of dimension three with four vertices, six edges, and four
sides. The base and extended topologies available in Shards are shown in Figure A.8.
The edge numbering is shown in Figure A.9 and the face numbering in Figure A.10.

22
 2 1

Figure A.3: Triangle edge (1-cell) numbering in Shards.

3 2 3 6 2 3 6 2

8
7 5 7 5

0 1 0 4 1 0 4 1

(a) (b) (c)

Figure A.4: (a) Base quadrilateral (Quadrilateral<4>), (b) eight-node ex-

tended quadrilateral (Quadrilateral<8>), and (c) nine-node extended quadrilateral
(Quadrilateral<9>) topology in Shards.

A.7 Pyramid

A pyramid is an object of dimension three with five vertices, eight edges, and five
sides. The base and extended topologies available in Shards are shown in Figure A.11.
The edge numbering is shown in Figure A.12 and the face numbering in Figure A.13.

A.8 Wedge

A wedge is an object of dimension three with six vertices, nine edges, and five sides.
The base and extended topologies available in Shards are shown in Figure A.14. The
edge numbering is shown in Figure A.15 and the face numbering in Figure A.16.

23
 2

3 1

Figure A.5: Quadrilateral edge (1-cell) numbering in Shards.

3
3 2
4 2

4 1

0 1 0

(a) (b)

Figure A.6: (a) Base pentagon (Pentagon<>) topology and (b) edge (1-cell) num-
bering in Shards.

A.9 Hexahedron

A hexahedron is an object of dimension three with eight vertices, twelve edges, and
six sides. The base and extended topologies available in Shards are shown in Figure
A.17. The edge numbering is shown in Figure A.18 and the face numbering in Figure
A.19.

24
 4 3 3

4 2

5 2

5 1

0 1 0

(a) (b)

Figure A.7: (a) Base hexagon (Hexagon<>) topology and (b) edge (1-cell) numbering
in Shards.

3 3

4 5
2 2
0 0 6

1 1

(a) (b)

7 9

8
6 2
0

4 5

(c)

Figure A.8: (a) Base tetrahedron (Tetrahedron<>), (b) eight-node ex-

tended tetrahedron (Tetrahedron<8>), and (e) ten-node extended tetrahedron
(Tetrahedron<10>) topology in Shards.

25
 3 5
4
2
0 1

Figure A.9: Tetrahedron edge (1-cell) numbering in Shards.

2
0 1

Figure A.10: Tetrahedron face (2-cell) numbering in Shards.

26
 4 4

12
11
9
10
3 3 7
8
2 2
0 0 6
5

1 1

(a) (b)

12
11
9
3 10
7
8
13 2
0 6
5

(c)

Figure A.11: (a) Base pyramid (Pyramid<>), (b) thirteen-node extended pyramid
(Pyramid<13>), and (c) fourteen-node extended pyramid (Pyramid<14>) topology
in Shards.

7
4 6
5

3 2

0 1

Figure A.12: Pyramid edge (1-cell) numbering in Shards.

27
 3 2
0 1
4

Figure A.13: Pyramid face (2-cell) numbering in Shards.

5 14 5
3 3
12 13

4 4 11
9

8
2 10 2
0 0
6 7

1 1

(a) (b)

14 5
3
12 13

4 11
9 17
15 16

8
10 2
0
6 7

(c)

Figure A.14: (a) Base wedge (Wedge<>), (b) fifteen-node extended wedge
(Wedge<15>), and (c) eighteen-node extended wedge (Wedge<18>) topology in
Shards.

28
 5
3 4

8
6

7
2
0 1

Figure A.15: Wedge edge (1-cell) numbering in Shards.

2
0 1

Figure A.16: Wedge face (2-cell) numbering in Shards.

29
 7 7
18
19
6 6
4 4
17
16
5 15 5
14
12
3 3 10
11 13
2 2
0 0
9
8
1 1

(a) (b)

7
18
19
6
22
4
17
16
15 5 26
23
20 14
12 24
25
3 10
11 13
21
2
0
9
8
1

(c)

Figure A.17: (a) Base hexahedron(Hexahedron<>), (b) twenty-node extended

hexahedron (Hexahedron<20>), and (c) twenty seven-node extended hexahedron
(Hexahedron<27>) topology in Shards.

7 6

4 5
11
10
8
3 2
9

0 1

Figure A.18: Hexahedron edge (1-cell) numbering in Shards.

30
 5
2
3
0 1

Figure A.19: Hexahedron face (2-cell) numbering in Shards.

31
DISTRIBUTION:

1 MS 1320
Kenneth Alvin, 1414
1 MS 1318
Roscoe Bartlett, 1411
2 MS 1320
Pavel Bochev, 1414
1 MS 1320
Scott Collis, 1414
2 MS 1541
Harold Edwards, 1541
1 MS 0316
Rob Hoekstra, 1437
1 MS 1318
Roger Pawlowski, 1416
1 MS 1318
Eric Phipps, 1416
2 MS 1318
Denis Ridzal, 1411
1 MS 1320
Andrew Salinger, 1414
1 MS 0316
John Shadid, 1437
1 MS 1318
Bart van Bloemen Waanders,
1411
2 MS 0612
Review & Approval Desk, 4916
2 MS 9018
Central Technical Files, 8944
2 MS 0899
Technical Library, 4536

32
v1.25